File size: 3,713 Bytes
a948150 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 | ---
license: mit
---
# 📡 TelecomTS: A Multi-Modal Telecom Dataset
**TelecomTS** is a large-scale, high-resolution, multi-modal dataset derived from a **5G telecommunications testbed**. It is the first public observability dataset to preserve **deanonymized** observability metrics with **absolute scale information**, encompassing by design various downstream tasks beyond forecasting such as **anomaly detection, root-cause analysis, and multi-modal reasoning**.
Observability data, particularly in telecommunications, differs fundamentally from conventional time series (e.g., weather, finance) by being:
- **Zero-inflated**
- **Highly stochastic and bursty**
- **Structurally noisy with minimal discernible temporal patterns**
These characteristics make TelecomTS a challenging benchmark for both language and time series foundation models. Our benchmarking experiments reveal that state-of-the-art models struggle to capture the abrupt, noisy, and high-variance patterns inherent in observability streams.
## 🚀 Key Features
- **32k Data Samples**
- **1M+ Observations** from a live 5G network
- **Multi-modal inputs**:
- Time-series KPIs across PHY, MAC, and network layers
- Environment descriptions and natural language QnA pairs
- **Absolute scale preserved** (no normalization/anonymization)
- **Downstream tasks supported**:
- 📈 Forecasting
- 🔎 Anomaly detection
- 🛠️ Root-cause analysis
- 🤖 Multi-modal question answering (time series + text)
- **Labels provided**: zone, application, mobility, congestion state, anomaly presence
## 📂 Dataset Structure
The main dataset consists of JSONL files containing chunked time series (128 timesteps each) along with multi-modal information. Each sample within the JSONL files includes:
- **Start_time / end_time** – temporal boundaries of the chunk
- **Sampling_rate** - number of timesteps per second
- **Description** – natural language summary of the network environment and time series behaviors
- **KPIs** – key performance indicator names and values
- **Anomalies** – existence, type, anomaly duration, affected KPIs, and troubleshooting tickets
- **Statistics** – mean, variance, and trends of the KPIs
- **Labels** – contextual metadata (zone, application, mobility, congestion, presence of anomalies)
- **QnA** – natural language reasoning tasks over the sample
Beyond the main dataset, each scenario also includes:
- **description.txt** – textual description of the network environment
- **metrics.csv** – raw observations for the scenario
## 🧪 Installation & Usage
### Install 🤗 Datasets
```bash
pip install datasets
```
### Load the Dataset
```python
from datasets import load_dataset
# Load the full dataset
dataset = load_dataset(
"AliMaatouk/TelecomTS",
data_files={"full": "**/chunked.jsonl"}
)
print(dataset)
```
### Inspect a Sample
```python
sample = dataset["full"][0]
print(sample.keys())
# dict_keys(['start_time', 'end_time', 'sampling_rate', 'KPIs', 'description', 'anomalies', 'statistics', 'labels', 'QnA'])
```
## Citation
You can find the paper with all details at https://arxiv.org/abs/2510.06063. Please cite it as follows:
```bib
@misc{feng2025telecomtsmultimodalobservabilitydataset,
title={TelecomTS: A Multi-Modal Observability Dataset for Time Series and Language Analysis},
author={Austin Feng and Andreas Varvarigos and Ioannis Panitsas and Daniela Fernandez and Jinbiao Wei and Yuwei Guo and Jialin Chen and Ali Maatouk and Leandros Tassiulas and Rex Ying},
year={2025},
eprint={2510.06063},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2510.06063},
}
``` |